Farid Fadaie is the cofounder and CEO of Viva AI, building AI tools for dental and healthcare operations.
Voice-agent bugs don't reproduce — they get reconstructed from logs. Decision-level logging, the trust boundary for LLM-written data, an anti-pattern bestiary from production, and why your code's second maintainer is an AI.
Conventional wisdom says a 50-location group is far ahead of the solo practice on AI. Often it's behind: growth by acquisition doesn't scale AI, it scales AI Sprawl. Scale is not maturity.
Conventional wisdom says big health systems will lead on AI. On the Maturity Model, the opposite is likely: when the bottleneck is deployment, not capability, the independent practice's small size is its advantage.
Healthcare AI is competing on the wrong variable. The most capable AI rarely wins — the one that ships does. Capability is a commodity; deployment is the moat.
Healthcare organizations plateau because no one owns the operating layer — the patient's journey across channels as one system. The durable question isn't who to hire; it's who owns AI transformation.
Healthcare organizations don't plateau at AI-Assisted because the technology isn't ready. They plateau because AI transformation is an organizational problem, not a technology one — and Stage 4 isn't something you buy, it's something you become.
As AI takes over the operational work, the question isn't what's left for people — it's which decisions must stay human. Execution scales; judgment doesn't.
After Operations-First AI comes the obvious question: how does an organization actually become AI-native? The Healthcare AI Maturity Model — Analog, Digital, AI-Assisted, AI-Native — is the map.
Health systems have the most AI money can buy and are the most stuck. The governance that makes clinical AI safe makes operational change impossible — which is why the biggest organizations plateau longest.
Healthcare doesn't need more AI tools — it needs an AI architecture: operations first, humans on the decisions, orchestration over automation, and production over demos.